(ed. This analysis was originally published in August 2013 and some of the comments below are from that original publication.)
One of the most famous Turkmen main carpet (MC) is also one of the most enigmatic and puzzling.
The ex-Ballard collection MC, now in the collection of the Metropolitan Museum of Art in New York
Were it not cut in half and rejoined perhaps a good part of that enigma would not exist. But since it is, there is really no way to know exactly how certain elements originally appeared, and the enigma goes on.
RK first saw this rug in the flesh sometime in the late 1970’s when we attended a private viewing at the Metropolitan Museum of Art at the invitation of our dear departed friend Charles Grant Ellis, and after several subsequent viewings we must admit our appreciation of its mystery and magic has not faded one bit.
The main obstacle to understand this MC remains in the fact no real analog has yet appeared; that said, there is one MC which is related enough to provide some important clues.
So-called multi-gol carpet sold at sothebys in 1999, and then gifted to the deYoung Museum in San Francisco by George Hecksher.
This relationship has not gone unnoticed and several years ago an article appeared in that rag hali with the rather unfounded premise the deYoung MC was “not Turkmen” but rather “Qizilbash”.
Written by an author whose previous efforts demonstrate he would be wiser remaining silent, this attempt was equally questionable. Especially considering it ends with this paragraph:
“ Thus when we first rejected the idea of the Hecksher(deYoung) carpet being Turkmen, we were in a sense quite wrong, for while it may not be a product of Central Asian Turkmen weavers(although the Gireyli might be a candidate), it is very likely to be from one of the northern Khorasan Qizilbash groups, perhaps Qajar, Afsher or even Bayat, all of whom were originally of Anatolian Turkmen origin.”
This passing the hat assumption does little to advance its understanding, or does it add any confidence to the already suspect ones the article tries to advance. In fact it negates them and returns the reader, and the confused author, back to square one.
So let’s now take a closer look at these two carpets, prove that relationship and how they fit into the larger Turkmen design and provenance panoply.
While what appears below is conjecture, some of our ideas are more concrete than others.
Let’s begin with those and work our way into to what might be called tripping into Turkmen fantasy-land.
The top to bottom cut the Ballard rug suffered has obviously caused severe dislocation of the original iconography. And no icon has been more destroyed than this.
Ballard rug’s most central icon, which after the cut has taken on a two-tree appearance.
Thanks to the excellent preservation of the deYoung rug, and the presence of the icon pictured below, we can pretty positively conclude the Ballard rug’s dislocated “tree-like” motif was originally the same – the key is both share blue-ground side panels, resembling paisley or boteh, filled with red chevrons.
Complex shield carpet-like icon, deYoung Museum MC
RK believes the Ballard rug to be the earlier of the two, which explains this icon’s simpler, less complex format.
But this is not the only icon these Turkmen MC share, and should readers question this relationship we suggest a careful comparison of the other gol.
Another truncated icon, actually a gol, from the Ballard rug with a related, complete one, from the deYoung.
Related gol; Left: Ballard, Right: De Young Museum
Again, there should be little to no doubt a relationship exists or, as important, the de Young version shows what the Ballard gol would have looked like prior to dissection.
And the more ‘complex’ form again the result of the deYoung’s carpet later origin.
Comparing the main borders of the two rugs also supports this contention, as the Ballard rug’s elegant and archaic version of the medallion/bracket/vine meander provides convincing proof it is earlier compared to the deYoung’s border of bi-colored trefoils, which are more frequently seen on trans-Caucasian and especially northwestern Persian weavings.
Although the author of the hali article provided not one piece of documentation to support Khorasan as the place of origin for the deYoung MC, this could have been a salient point to cite.
In the same vein would be the already mentioned similarity the “Shield” carpet icon has with that group of weavings, as some of those carpets can quite probably trace their origins to workshops in that general vicinity.
An additional, albeit minor, note is the deYoung MC’s version of the two identical secondary border that are sometimes seen in Classic period MC, not in Archaic period ones by any means.
The Ballard rug’s simple barber-pole minors are perfectly in keeping what RK considers to be an earlier, Archaic period, format and it is difficult not to notice how they provide artistic balance to the complex white ground main border.
Worth another small mention is the Ballard MC’s main border that is, again in our perception, the archetype for the later ‘boat’-type main border seen on some righteously old eagle-group and other Yomud MC.
But one icon, gol, these two MCs do not seem to share is this
However, the deYoung’s carpet weaver, or designer, must have somewhat known it, as this later version implies this was the case.
Once more the Ballard MC’s version, which is far more sophisticated, but not necessarily more ‘complex’, demonstrates what we see as the archetype. Note Bene: Complexity does not archaic make; rather a sophisticated, unified pattern is always more likely to be early than a overly complex, accreted, one.
And this is the case here, with the deYoung’s. It is clearly a descendant copy, where some essential elements have been lost and simplified, the whole form accreted by the addition of foreign elements.
Need we cite the pair of yellow-brackets at each side which, unlike the Ballard’s non-symmetric form, have rendered the deYoung’s gol symmetric.
For us there is little doubt the Ballard gol was involved in the mix-master that produced the deYoung’s. As proof notice its nascent brackets at each side and are those three pair of 'wing-like' devices at bottom of the field of the deYoung MC the vestigial remains of the orange colored diagonal stripes seen in its gol?
Another gol missing from the deYoung MC is the “C” gol, and RK finds this quite interesting.
Obviously the deYoung rug is a Turkmen-type rug; related, but one very far removed from the mainstream of Turkmen weaving traditions.
The historic movement of certain Turkmen groups, sometimes over great distances, is now becoming more known. Caused by a myriad of political, economic and environmental stimulus, these moves brought the adoption of alien designs into the Turkmen weaving tradition, as well as having the reverse effect with alien weaving vocabularies being influenced by the Turkmen.
This is the best explanation we can offer to explain just how a rug like the deYoung’s happened to be produced; but neither we, nor anyone else, can definitively state it was the result of the former, or the latter, paradigm.
Was this somewhat beautiful, but strange, MC produced for a Turkmen audience, or client?
This is likewise an intriguing question; however, it too unanswerable.
In conclusion we do not believe the weaver was, in a strict sense, Turkmen, the materials and coloration stand to support this idea. And although we cannot prove this, it seems far more likely than thinking the deYoung multi-gol MC deserves to be called a Turkmen rug.
Another item in this tack is the absence of the “C” gol icon, which is there for all to see in the Ballard MC.
“C” gol, Ballard MC
This absence, for us, is a notable indication of its non-Turkmen origin, as all the other gol are basically found on other genre of weaving, while the “C” gol is, to the best of our knowledge, seen only, and exclusively, on Turkmen MC.
Let’s take another look at perhaps the most unusual and strange gol iconography both rugs share.
Two details: Right: deYoung MC; Left: Ballard MC
Were this large, weirdly configured, gol not on an archaic example like the Ballard MC, RK might be moved to place it quite a bit farther down on the dating continuum.
But it is there, begging us to explain what it is all about.
It is always easier to see where things are going, at least in an art historic sense than where they came from, and such is the case here.
Here this detail from a unique early Classic period MC bears some attention, for between its two row of “C” type gol there is what we believe to be the kepse gol prototype, as this is earliest kepse gol we have ever seen.
Detail, double or dual kepse gol
Here is the complete MC
RK realizes, as they say, it’s a long way to Tipperary to imply this ‘double kepse’ is related to the gol both the Ballard and deYoung MC’s display.
However that said, here’s our rational.
We posit they both developed from a presently unidentified, unknown earlier gol, and as such represent two branches off that stock.
By twisting and rotating the individual elements of the kepse gol it is possible to recreate most of those in the Ballard/deYoung MC gol. And the unique double aspect of the kepse so well echoes what appears in the ‘doubled’ deYoung and Ballard gol we couldn’t resist.
This is just guesswork, what RK referred to as tripping in Turkmen fantasy-land.
And hypothetical as it might be, it is the only explanation how, or from where, such a gol might have developed.
Back to slightly firmer ground RK would like to add a short bit of commentary on another Ballard MC gol eccentricity
Unique gol with Timurid/Safavid style metalwork element, a detail of which appears on the right
This seemingly insignificant squiggle repeated eight times, four single and four double, is actually quite significant, as we know of no other Turkmen weaving where it so clearly appears or is even implied.
And while RK is not the first to mention, or believe, one root of Turkmen gol iconography is planted within the mythical Timurid weaving tradition, we are now the first to provide any clue of documentary evidence.
Unfortunately no Timurid weavings exist, they are only known from illustrations in miniature paintings of the period.
Needless to say this is a major obstacle, but there are some examples of metalwork, which can be assigned to Timurid hands.
Here is a Timurid brass and silver inlaid candlestick, dated to “before mid-15th century” illustrated in Pope “Survey of Persian Art”.
And a detail showing the Ballard MC scrolling leaf beside one from the Timurid candlestick.
Left: detail Timurid candlestick; Right: detail Ballard MC
This iconic squiggle is probably a transitional stage of the later Ottoman, and Safavid period, saz-leaf.
Ottoman tile with saz-leaf and rosette pattern; circa 1545-1555, Metropolitan Museum of Art, NY
It also bears relation with so-called strapwork elements appearing on some groups of Safavid period Persian carpets.
Discovering this squiggle on the Ballard gol could very well link it with as yet unknown and unidentified Timurid/Turkmen gol iconography.
In any event, it was no accident that put it there; it is but one additional proof for the unique and archetypal status the Ballard MC maintains.
In conclusion we can add one more anecdotal reference why we believe the Ballard MC to be a Turkmen product and the deYoung MC not.
small tertiary tree icon, Ballard MC
Icons like this simple tree and the five spot, or x, which are scattered on its field are more than a gentle reminders of other early Turkmen MC, like this Saryk Timurchin gol, where almost exact same ones appear.
Detail of tertiary simple tree and five spot, x, icons; archetypal Saryk Timurchin gol MC; RK collection
In comparison it is very interesting to note the deYoung MC tertiary elements are distinctly non-Turkmen in character. And this minor point might have be far more significant than it might seem at first glance.
Well, RK believes a weaver was far more likely to feel ‘free’ in their use and placement than with major and minor gol, or a main border. And that’s why their presence on the Ballard MC has a Turkmen character, and those on the deYoung MC a non-Turkmen character.
A few questions to wrap this up.
Does the deYoung MC provide strong analogs to understanding the Ballard MC?
We believe the answer is Yes.
Is the Ballard MC a Turkmen product?
We believe the answer is Yes.
Is the deYoung MC a Turkmen product?
We believe the jury is still out on this one. But from the available evidence we would presently have to say No.